JP2009507261A - Audio signal playback system - Google Patents

Abstract

An apparatus and a method for reproducing and transmitting a first audio signal (4) and a second audio signal (5), a coding means (10) for a second audio signal (5), and a first audio signal ( 4) and the coded second audio signal (5e) are combined to produce a synthesized audio signal (6) that can be reproduced by the first audio reproduction means (8), producer (1), first audio reproduction It comprises a transmitter (2) having means (8), a receiver (16) for receiving a synthesized speech signal (6) and a receiver (3) having an extracting means (17).
[Selection] Figure 1

Description

  The present invention is directed to a system that plays audio and transmits audio signals wirelessly, and a “speaking” toy that reproduces audio signals that are inaudible and invisible wirelessly transmitted on television. Related to the application.

  Talking toys with similar objectives of the present invention are described in WO0169572 [Creator Ltd.], DE 1520586 [Siemens AG], US 2004/082255 [PSL et al. L. Fong et al.)], US 6238262 [Technology Australia Pty Ltd.], US 5191615 [The Drummer Group], US 4846693 (Smith Engineers). Smith Engineering)], U.S. Pat. No. 4,840,602 (Coleco Industries, Inc.) Which is incorporated herein by reference.

  US Pat. No. 5,191,615 describes a toy (doll) that is closest to the present invention and that can play a soundtrack received from outside and can be controlled by external signals. This interactive toy can move and play audio in sync with the video program's soundtrack, so that the toy appears to interact with the video program by responding to the soundtrack .

  The US 5916615 system / toy uses the two audio channels of a stereo system, either a television transmission signal or a recording on a recording medium (cassette or DVD). The first audio channel carries the audio of the video program intended to be played on the television, and the second audio channel plays the audio signal (played by the toy drive system) intended to be played by the toy. (With a coded signal).

  When the television signal is received by the television or when the recording medium is reproduced, the first audio channel is separated by the splitter and directly reproduced by the television, and the second audio channel is transferred to the toy. By demultiplexing, the sound played by the speaker on the toy is separated from the signal encoding the motion indication either before transfer or after transfer (ie in the toy).

The transmission between the splitter and the toy is preferably wireless so that the toy can move freely. US Pat. No. 5,191,615 discloses the use of frequency modulation (FM) radio links or infrared links.
In this type of application, the infrared link is too directional and weak enough to keep up with the movement of children moving around with toys.
The radio link has a problem of administrative approval. For example, in Europe, there are very strict regulations, and at least development in conformity with the regulations of the government that assigns frequencies is essential.
Another drawback common to these two transmission modes is that it requires the use of a television equipped with a transmitter specially adapted for infrared or radio transmission.

  The present invention remedies these drawbacks, and when the transmitter / television reproduces the first audio signal, the second audio signal is coded so that it cannot be heard and combined with the first audio signal. It is possible to provide an advantageous transmission form of being reproduced. However, this audio signal can be received by the receiving means in the receiver / toy. The receiver extracts the second audio signal, decodes it, and plays it back so that it can be heard.

The present invention is a method for generating a synthesized voice signal from a first voice signal and a second voice signal, and includes the following steps.
Encoding the second audio signal into a coded second audio signal that can be reproduced in a manner inaudible to the human ear by the audio reproduction means;
The synthesized voice that can be reproduced by the voice reproduction means by combining the first voice signal and the coded second voice signal in such a way that the first voice signal can be heard by the human ear and the second voice signal cannot be heard by the human ear. The voice combining stage to be a signal.

  This synthesized audio signal is compatible with general audio signals and is preferably transmitted, recorded and reproduced by existing means without modification. When reproduced by standard audio reproduction means, the first audio signal is reproduced so that it can be heard, but the second audio signal is transmitted in such a way that it cannot be heard by the human ear. Thus, the transmitter with the standard first reproduction means reproduces the first audio signal so that it can be heard, and at the same time, without additional special means, transmits the second audio signal wirelessly to the receiver so that it cannot be heard. it can. The receiver can reproduce the second audio signal so that it can be heard by the second audio reproduction means. In this way, it is possible to give the impression that the two playback means are interacting.

According to another feature of the invention, the encoding step comprises a frequency shift step of shifting the frequency of the second audio signal to a high frequency that is not audible to humans in the reproduction frequency spectrum of the audio reproduction means.
In this way, the second audio signal is in a frequency spectrum that can be reproduced by standard audio reproduction means, but becomes inaudible to human ears, is protected wirelessly, and does not require additional transmission means. Can be played.
In the frequency shift, the first audio signal and the second audio signal do not overlap with each other so that the two can be coupled.

  According to the invention, the coding stage comprises a band reduction stage of the second audio signal. This reduction operation facilitates the join operation.

  In the present invention, the encoding stage may further include an encryption stage. This stage requires decryption of the code, but allows the second audio signal to be heard at its destination.

According to the invention, the encryption step comprises the step of removing the positive part of the frequency spectrum of the second audio signal and leaving only the negative part.
Even if the embodiment is perceived by a person with ultra-sensitive hearing, its contents cannot be understood, and it has two benefits of facilitating the speech combining step with a more adaptive spectral shape. is there.

  According to another feature of the invention, the speech combining stage combines the first speech signal and the coded second speech signal. This step has the advantage of being easy to implement.

  According to another feature of the invention, the speech combining stage provides a first gain to the first speech signal and a second to the second speech signal before combining the first speech signal and the coded second speech signal. Giving gain.

  According to the present invention, the second gain is preferably lower than the first gain. This has an advantage that the signal masked with the code (second audio signal) is placed under a general auditory curve to improve the concealment condition.

According to another advantageous characteristic of the invention, the speech combining step comprises the step of removing the high frequency part of the frequency spectrum of the first sound signal before combining and retaining only the low frequency part.
By the operation of removing the high frequency part, the first audio signal is virtually inaudible, and a part of the reproduction frequency spectrum is released, and the audio combining operation becomes very easy.

  According to another feature of the invention, the combining step further comprises the step of applying a command signal at a frequency Fcmd.

  In the present invention, the frequency Fcmd is higher than that of the coded second audio signal, lower than 15625 Hz for the PAL signal, and lower than 15734 Hz for the NTSC signal.

  According to another feature of the invention, the command signal is a binary command coded with pulse width modulation at 31.25 bits / second.

  According to another optional feature, the synthesized speech signal is transmitted with an identification code identifying at least the intended receiver. This feature is that communication is performed between the transmitter and a plurality of receivers to which the respective audio signals are assigned.

  Also, according to another optional feature, an identification code with the word “Broadcast” is given to all receivers. This identification code of << Broadcast >> should be compatible with all receivers and be reproducible by all receivers.

  The present invention further relates to a synthesized speech signal obtained by the method according to the above embodiment. The present invention also extends to a recording medium for a synthesized audio signal.

  The present invention further relates to a producer that creates a synthesized speech signal.

  The present invention further relates to a method of reproducing a first audio signal and wirelessly transmitting a second audio signal, wherein the synthesized audio signal according to the above-described embodiment is generated from the first audio signal, the second audio signal, and possibly a command signal. There are stages of generation and voice reproduction of the synthesized voice signal by the voice reproduction means.

  The present invention further relates to a method for receiving a synthesized speech signal, which receives the synthesized speech signal according to the above-described embodiment, extracts a coded second speech signal from the synthesized speech signal, and decodes the coded second speech signal. The second audio signal is reconstructed and the second audio signal is reproduced.

  According to another feature of the invention, the speech receiving method includes the steps of extracting a command signal from the synthesized speech signal, demodulating the command signal, and reconstructing a binary command.

  According to another feature of the present invention, the decoding stage or the demodulating stage has a stage that is symmetric with respect to the coding or modulation stage in the synthesized speech signal creation method of the above embodiment.

  According to another characteristic of the invention, the decryption means comprises a << mirror >> transform that reconstructs the positive part of the second audio signal from the negative part of the frequency spectrum of the second audio signal.

  According to another feature of the invention, a gain / gain control step is applied to the second audio signal prior to the audio reproduction step. The audio reception signal varies depending on the distance between the transmitter and the receiver, but this gain / gain control step can reproduce the second audio signal at the same volume independent of the intensity of the audio reception signal.

  According to another feature of the invention, the audio reception signal has a noise removal stage of the second audio signal based on the estimation of the noise spectrum before reproduction.

  According to another feature of the invention, the noise spectrum is regularly measured during quiet periods or treated as blank noise.

  According to another feature of the invention, the receiver has its own identification code, and at least the identification code identifying the intended receiver is extracted from the synthesized speech signal and extracted from its own identification code and the synthesized speech signal. Only when the code is compatible, the receiver plays the second audio signal.

  According to another characteristic of the invention, the identification code << broadcast >> corresponds to all receivers, the identification code << broadcast >> is extracted and the second audio signal is reproduced.

  The present invention further relates to a receiver that can receive a synthesized speech signal and implement the speech receiving method according to the above-described embodiment.

The present invention further relates to a toy having the above receiver.
This allows a “speaking” toy that appears to be interacting with the audiovisual program, being sent wirelessly to audiovisual means such as television so that the “voice” is not known.

  The present invention further generates the synthesized voice from the first voice signal and the second voice signal and transmits the synthesized voice to the transmitter or the first voice playback device, plays back the first voice signal, and plays the second voice. The present invention relates to an audio signal reproduction system having a transmitter or a first audio player according to the above embodiment for transmitting a signal wirelessly, and a receiver or a second audio player according to the above embodiment for extracting a second audio signal and decoding and reproducing it. Is.

  The features, details and advantages of the present invention will become more apparent with reference to the following detailed description and accompanying drawings.

In the figures, the same reference numerals indicate the same or similar items. Reference letters and reference numerals are used for different forms of the same object. For example, reference numerals 5a-5e indicate the second audio signal 5 in the various forms (reduction, encryption, coding) according to the various processes of the present invention as it is transmitted.
5 to 16 show audio signals. In either case, the frequency spectrum shows the frequency in Hz or kHz on the horizontal axis and the volume in dB on the vertical axis.

  FIG. 1 shows one embodiment of the system of the present invention. The system consists of a producer 1, a transmitter-2, and at least one receiver 3.

The producer 1 receives the first audio signal 4 and the second audio signal 5. At this stage, the first audio signal 4 is in the basic format 4a shown in FIG. At this stage, the second audio signal 5 is in the basic format 5 a shown in FIG. 8 and is substantially the same as the basic format 4 a of the first audio signal 4. The producer 1 is a signal processing unit, and generates a synthesized audio signal 6 from the first audio signal 4 and the second audio signal 5 by a method described later.
The synthesized audio signal 6 is transmitted to transmitter-2 by direct connection 30, playback connection 31, recording medium 24 or equivalent means well known to those skilled in the art.

  The transmitter 2 is a first audio reproducing device, and the first audio reproducing means 8 reproduces the synthesized audio signal 6, reproduces the first audio signal 4 in a form that can be heard by the human ear 7, and the second audio signal. 5 is reproduced in a coded format 5e that cannot be heard by the human ear 7.

  One or a plurality of receivers 3 are provided in the second sound reproducing device, and the receiver 3 is placed within the sound reproduction range of the transmitter 2 and receives a portion of the synthesized sound signal 6 including the coded second sound signal 5e. The receiver 3 extracts the encoded second audio signal 5e, decodes it, and reconstructs it into the basic format 5a of the second audio signal. Next, the receiver 3 or the second audio reproduction device uses the second audio reproduction means 9 Thus, the second audio signal 5 that can be heard by the human ear 7 is reproduced.

  FIG. 2 is a detail of one embodiment of the producer 1. The producer 1 has the encoding means 10, encodes the second audio signal 5 and converts it into the encoded second audio signal 5 e. Further, a combining unit 11 for combining or multiplexing is provided, and the first audio signal 4 and the encoded second audio signal 5e are combined to form a synthesized audio signal 6.

  When reproducing the synthesized speech signal 6, the coding means 10 shifts the frequency spectrum of the second speech signal 5 to a frequency higher than the frequency audible to the human ear 7 in order to make the second speech 5 inaudible. It is preferable to have one frequency shift means 12.

  FIG. 12 shows the coded audio signal 5e shifted by Δ amount in the high frequency direction of the frequency spectrum. In FIG. 12, a curve C indicates a lower limit of a frequency spectrum that can be normally heard by the human ear 7 and a general auditory curve C. However, the frequency shift Δ is set so that the coded second audio signal 5e thus obtained is within a frequency spectrum that can be reproduced by the conventional first audio reproduction means 8. The first frequency shift means 12 is an essential means for the encoding means 10. The first frequency shift means 12 can be implemented by, for example, a frequency multiplex device (multiplicator de frequency).

  In order to limit the frequency spectrum of the reproduced sound, it is preferable that the encoding unit 10 further includes a band reducing unit 13 that further reduces the band of the second audio signal 5. The method or means for reducing the frequency spectrum is aimed at reducing the bandwidth of the frequency spectrum used by the signal. This reduction can be performed by compressing the frequency spectrum or selectively and deliberately deleting the frequency spectrum so that only a limited band remains. Still, the rest is in good condition to recover. For example, if the frequency is about 3 kHz including the most frequently used frequency, it is possible to recover the initial audio signal. With such a reduction method, the basic format voice 5a shown in FIG. 8 can be converted into the reduced spectrum voice 5b shown in FIG.

  In one embodiment, the frequency spectrum is reduced with a filter that passes a low band at a cutoff frequency Fc = 3 kHz, modulated to a carrier frequency Fp = 14.25 kHz with BLU modulation, and then the band between Fp and Fp + Fc is Through the filter that passes, the audio signal finally has a reduced spectrum of 3 kHz. The encryption step using the negative band, described below, simultaneously reduces the encrypted signal with band [11.25; 14.25] kHz through a filter that passes through the band between Fp-Fc and Fp. Is advantageous.

  The encoding means 10 can further comprise an encryption means 14 for the second audio signal 5. The encrypting means 14 prevents the contents of the second audio signal 5 from being understood, so that only a legitimate receiver can decrypt or encrypt or copy the signal.

  FIG. 10 shows a reduced spectrum speech signal 5b having a full frequency spectrum composed of a negative portion 5c and a positive portion 5d of the speech signal. The negative part 5c of the frequency spectrum is always present in the audio signal and is symmetrical to the positive part 5d with respect to the vertical axis. Since the negative part 5c has no additional information, it is generally omitted to express the frequency. One example of the encryption method performed by the encryption means 14 is to remove the positive part 5d of the second audio signal 5 and retain only the negative part 5c of the frequency spectrum. Such encryption can be performed by the encryption means 14 using a low-pass filter that holds only the negative portion 5c of the frequency spectrum. Encryption can also be performed by applying a “mirror” symmetry process to only the positive part 5d of the frequency spectrum.

  There are two purposes for encryption. The first purpose is encryption itself, so that even if a person with ultra-sensitive hearing senses an encrypted signal, the contents cannot be understood. The second purpose is to create a new frequency spectrum shape to facilitate the speech combining stage, and as shown in FIG. 12, the new shape frequency spectrum of the negative part 5c of the speech signal is coded second speech. Shifted to signal 5e and easily hidden under the auditory curve C or hidden under the auditory curve C with a small shift Δ.

  The combining means 11 is to construct the synthesized speech signal 6 in the first audio signal 4 in one format and the second audio signal 5 in one format, for example, the coded second audio signal 5e. As shown in FIG. 13, each of the first audio signal 4 and the second audio signal 5 is a reproduction frequency spectrum that can be reproduced by the first audio reproduction means 8 although the frequency spectra of the two audio signals do not completely overlap. Like that. Therefore, the coupling which is frequency multiplexing is preferably performed by the coupling means 11 having an adder 15 which adds the first audio signal 4 and preferably the second audio signal 5 in the encoded form 5e.

  Before the addition, it is preferable to apply a first gain to the first audio signal 4 and a second gain to the second audio signal 5. The gain can compensate for attenuation of one of the audio signals either before or after. In order to hide the coded second audio signal 5e below the general auditory curve C, the second gain is preferably lower than the first gain.

  Prior to performing the combining operation, the first audio signal 4 is preferably freed from high frequency parts by modifying the frequency spectrum. 6 and 7 illustrate this frequency spectrum correction. This operation removes the high frequency part 4c of the frequency spectrum of the first audio signal 4 and makes only the low frequency part 4b. The frequency F1 is selected so as to determine the upper limit of the frequency spectrum and separate the low-frequency portion 4b from the high-frequency portion 4c. By selecting the value of F1, the quality degradation of the first audio signal 4, the “space” available for the second audio signal 5, and the risk of interference between the first audio signal 4 and the second audio signal 5 Is determined. This operation can be satisfactorily performed by the low-pass filter 33 having the natural frequency F1. The low-pass filter 33 may be incorporated in the coupling means 11 or may be incorporated by another means on the input side of the coupling means 11 as shown in FIG. As shown in FIG. 7, the low-frequency portion 4b of the first audio signal is advantageously combined with the second audio signal 5 in the encoded form 5e.

  As shown in FIG. 13, the frequency spectrum portion vacated by removing the high frequency portion 4c does not overlap the two audio signals and does not overlap the bandwidth of the reproduction frequency spectrum that can be reproduced by the conventional reproduction means. This is used to arrange the second audio signal 5 in the encoding format 5e. The frequency F1 is selected such that the coded second audio signal 5e is in the reproduction frequency spectrum. If the second audio signal has been reduced in advance, the high frequency portion 4c removed from the basic format 4a is preferably narrow. Furthermore, this region of the frequency spectrum corresponding to high frequencies is not very important in terms of the sound quality perceived by the human ear 7. For this reason, removing the high frequency portion 4c of the first audio signal 4 and using only the low frequency portion 4b only slightly changes the first audio signal 4 sensed by the human ear 7, and the second audio This makes it very easy to combine the signal 5 and the first audio signal 4.

  In a particular embodiment, an identification code that can be identified by at least a given receiver 3 can be added to the synthesized speech signal 6. In this embodiment, the system includes at least a producer 1, at least a transmitter 2, and a plurality of receivers 3. Each of the plurality of receivers 3 is identified by its own identification code. The identification code may be different for each receiver 3 or may be common to a plurality of receivers 3 in one “family” intended to reproduce the same second audio signal 5. When the second audio signal 5 is generated by the producer 1 and transmitted by the transmitter 2, the latter audio signal has an identification code corresponding to the receiver in the synthesized audio signal 6 before the second audio signal 5 is reproduced. Added to. The receiver 3 in which the transmitted identification code does not correspond to its own identification code receives the second audio signal 5 but does not reproduce it.

  In another optional embodiment, it is limited by an identification code of the word “Broadcast”. This identification code is not assigned to the specified receiver 3, but is used only by being inserted into the synthesized speech signal 6. Regardless of whether or not the identification code is correct, all receivers 3 can recognize the identification code corresponding to the same receiver 3, and the receiver 3 can reproduce the second audio signal 5 having the << broadcast >> code. It is easy for a person skilled in the art to generalize what has just been shown, using identification codes to subset into a plurality of receivers 3, for example using addressing or masking techniques.

  The identification code can be inserted into the synthesized speech signal 6 by any means known to those skilled in the art. In one embodiment, this insertion can be done by combining the third audio signal forming the command signal 35. FIG. 15 shows the frequency spectrum of such a command signal 35. As will be described below, this signal is coded into a binary command and is in the form of a third audio signal. As shown in FIG. 15, this signal is a limited frequency spectrum having a high frequency Fcmd, and can be combined with the first audio signal 4 and the second audio signal 5 to form a synthesized audio signal 6. The frequency Fcmd of the command signal 35 is determined to be higher than the high frequency of the frequency spectrum of the coded second audio signal 5e. However, the frequency Fcmd is determined to be lower than the frequency emanating from the cathode ray tube, that is, 15625 Hz in the PAL mode and 15734 Hz in the NTSC mode. The command signal 35 is then combined with the first audio signal 4 and the encoded second audio signal 5e by frequency multiplexing to form a synthesized audio signal 6. FIG. 16 shows such a synthesized speech signal 6. As described above, the synthesized audio signal 6 is an audio signal that is completely included in the reproduction frequency spectrum of the conventional audio reproducing apparatus, like the synthesized audio signal 6 described so far. Accordingly, a playback device such as a speaker completely reproduces the components of the second audio signal 5e and the command signal 35 in a manner that cannot be heard by human ears, but in a manner that can be used by a receiver as described below. In all other respects, the synthesized speech signal 6 must be understood as a signal resulting from the combination of two or three signals.

  The binary command is suitable for encoding the command signal 35 by pulse width modulation. This modulation is described in the AES-42 standard. If Nb is an integer that determines the modulation time referred to as the “bit time”, F′p is the frequency of the modulated carrier, Fe is the frequency of the signal sample, and g is the variable power parameter, bit 0 is the pulse v (n) = G. cos (2nп (F'p / Fe)) where n = 0. . Nb / 4-I, v (n) = 0 is n == Nb. . Nb-I, bit 1 is pulse v (n) = g. cos (2nп (F'p / Fe)) where n = 0. . 3 (Nb / 4-1) and v (n) = 0 is n = 3Nb / 4. . It is encoded as Nb-1. Therefore, by selecting F′p = 14.625 kHz together with Fe = 48 kHz and Nb = 1536, information of 31.25 bits / s can be obtained.

  With reference to FIG. 2, the synthesized voice signal 6 produced by the producer 1 is then transferred to the transmitter 2. At the output of the coupling means 11, the producer 1 selects the direct connection 30, the emmetr 26 for playback connection 31, or the recording means 32 to the recording medium 24 when the synthesized audio signal is not intended for later use. Take it. The regenerative connection 31 can be any connection method known to those skilled in the art and can be implemented in a cable or wireless, analog or digital type. Examples include radio and television broadcasts, or transmission by FM playback connected via USB or ADSL. As the recording medium 24, for example, a well-known audio cassette, video cassette (such as VHS), DVD, CD, CD-ROM, or other equivalent medium capable of recording analog or digital audio signals can be used. The synthesized speech signal 6 is transmitted to the transmitter 2 in that one format.

  FIG. 3 shows details of the transmitter 2. At the entrance of the transmitter 2, there is a response means for receiving the synthesized voice signal 6 generated by the producer 1. This response means is a receiver 27 that can correspond to the direct connection 30, the emmetter 26 that is the reproduction connection 31, or a reading means 25 that can read the recording medium 24.

In another embodiment, the producer 1 is integrated into the transmitter 2.
The transmitter 2 has a first reproduction function for reproducing the first audio signal 4. In order to implement this, the first sound reproduction means 8 is essential. The first sound reproduction means 8 usually has at least one speaker or other similar device. Furthermore, it is preferable to have a processing means 29 having elements necessary for forming an audio signal before reproducing the audio. Here, the processing means 29 has means necessary for reading the recording medium 24 and converting the signal into a format that can be reproduced by the first sound reproducing means 8. These means are well known to those skilled in the art and are prior art and do not form part of the present invention.

  It should be noted that these elements are conventional and need not be changed in order to practice the present invention. The synthesized speech signal 6 has all the attributes of the basic format 4a (FIG. 5) or 5a (FIG. 8) of the conventional speech signal. Therefore, the purpose of reproducing sound can be achieved by the conventional first sound reproducing means 8. Thus, the transmitter 2 does not need to be modified according to the present invention. All existing audio playback devices can be used with the present invention. Cassette players, televisions, audio or video audio playback devices can be used. It is sufficient to input the input audio signal to the audiovisual means and change it to the synthesized audio signal 6. The conventional television can reproduce the synthesized audio signal 6 by using the synthesized audio signal 6 instead of the audio signal received from the television broadcast or the cable network. Similarly, the DVD player can record and reproduce the synthesized audio signal 6 on a DVD instead of a sound track.

When reproduction is performed by the first audio reproduction means 8 of the transmitter 2, all components of the synthesized audio signal 6 as shown in FIG. 13, that is, the component 4 b derived from the first audio signal 4 and the component 5 e derived from the second audio signal 5. In some cases, even components derived from the command signal 35 are reproduced. However, in consideration of the general auditory curve C shown in FIG. 13, only the portion located in the upper region of the curve C can be heard and recognized by the human ear 7. Most of the first audio signal 4 has a low frequency portion 4b and is above the curve C and is heard by the human ear 7. Further, the portion above the curve C is substantially the same as only the first audio signal 4 in the form of the basic format 4a (FIG. 5). As a result, the human ear 7 hears substantially the same sound as the first sound reproducing means 8 reproduces only the first sound signal 4 of the basic form 4a. The human ear 7 sounds as if the transmitter 2 had reproduced the first audio signal 4. For this reason, the transmitter 2 is also referred to as a first audio playback device.
The second audio signal 5 is in the encoded form 5e, and is located completely below the general auditory curve C together with the command signal 35 and is usually not recognized by the human ear 7 and remains inaudible. It is.

  However, in the synthesized voice signal 6, the portion that is derived from the second voice signal 5 and coded in the coded second voice signal 5 e that cannot be heard by humans, and in some cases, the command signal 35 is reproduced well. Should be noted.

  FIG. 4 shows details of an embodiment of the receiver 3 and the second sound reproducing device according to the present invention. The receiver 3 receives the synthesized speech signal 6, extracts the second speech signal 5 in possibly coded form 5e, and extracts the command signal 35. For this purpose, the receiver 3 has a receiving means 16 for the synthesized speech signal 6 and an extracting means 17 for extracting the second speech signals 5 and 5e coupled to the synthesized speech signal 6 and the command signal 35. For this purpose, the receiving means 16 has a band that can cover the frequency spectrum part of the synthesized speech signal including at least the second speech signals 5, 5 e and the command signal 35. It can be pointed out that the receiving means 16 need not have a spectral portion corresponding to the first audio signal 4. On the contrary, the receiving means 16, unlike the human ear 7, can receive a high frequency spectral portion of the synthesized speech signal 6 having the second speech signals 5, 5 e and the command signal 35. It is a feature.

  In the first embodiment, the receiving means 16 receives a part of the frequency spectrum of the synthesized speech signal 6 including a portion derived from the second speech signals 5 and 5e and a portion derived from the command signal 35. The extraction means 17 then separates only useful parts from the received signal. Therefore, the extraction means 17 is typically a bandpass filter. FIG. 14 shows the extraction stage. In the second audio signal 5 which is the encoded second audio signal 5e, only the frequency spectrum portion between the two frequencies F2 and F3 is extracted from the received frequency spectrum. Similarly, although not shown here, only the frequency spectrum portion between two encrypted frequencies is extracted from the command signal 35.

  In another second embodiment, the extraction means 17 has a band corresponding to the spectral portion between the two frequencies F2 and F3 together with the reception means 16, and the reception means 16 The second audio signal 5e and the command signal 35 can be related to each other. The receiving means 16 is typically a microphone.

  The receiver 3 further has decoding means 18. The decoding means 18 can decode the encoded second audio signal 5e, and reconstructs the second audio signal 5 into its basic format 5a. The receiver 3 further has second sound reproduction means 9 for reproducing the second sound signal basic format 5a so as to be heard.

  Those skilled in the art can easily infer from the present invention that various processing forms (encoding, encryption, reduction, transmission, acceptance, etc.) applied to the second audio signal 5 can be executed in analog or digital form. Will. Similarly, no matter what technique is adopted, it is clear that the processing time is very short and can be ignored. The second audio signal 5 is transmitted from the transmitter 2 to the receiver 3 for reproduction by the second audio reproduction means 9 and can be reproduced substantially simultaneously in real time.

  Those skilled in the art will realize that the two audio signals 4 and 5 can be used simultaneously by performing frequency separation of the first audio signal 4 and the second audio signal 5 in the synthesized audio signal 6. Therefore, the receiver / toy can “speak” or “sing” until the transmitter / television plays its audio signal, and at the same time without any order or arrangement of who speaks to who.

  Although the decoding means 18 can be implemented in various forms, it is preferably performed by a symmetric operation corresponding to the encoding means 10, and the decoding operation is performed in the reverse order to the encoding operation in the encoding means 10.

  The frequency shift stage is indispensable for encoding, and the decoding means 18 requires the second frequency shift means 21 for the encoded second audio signal 5e. The first frequency shift means 12 performs Δ frequency shift in the high frequency direction, while the second frequency shift means 21 similarly performs Δ frequency shift in the low frequency direction. This stage reconstructs the audio signal so that it can be heard. The second frequency shift means 21 is preferably a frequency demultiplexer so that the first frequency shift means 12 can preferably be executed by a frequency multiplexer. frequency). In the decoding means 18 and other related stages, there are as many other elements as there are symmetrical means in the encoding means 10, which can be arbitrarily selected.

  Therefore, when the encoding means 10 has the bandwidth reduction means 13, the decoding means 18 has a bandwidth recovery means 20 that is the reverse of the bandwidth reduction means 13 if necessary. When having the encryption means 14, the decoding means 18 has an encryption / decryption means 19 that is the reverse of the encryption means 14 described above.

  When the encryption means 14 of the encoding means 10 removes the positive part 5d of the frequency spectrum of the second audio signal 5 and retains only the negative part 5c, the decryption means 19 uses the positive part 5d and the negative part 5c. The reconstructing means 22 for the reduced spectrum speech signal 5b having the entire frequency spectrum is provided. Such reconstruction means 22 is, for example, << mirror >> means that can reconstruct the positive part 5d from the negative part 5c in consideration of symmetry with respect to the vertical axis.

  Those skilled in the art will recognize that the band reduction means 13 / band recovery means 20, the frequency shift means 12/21, the encoding / decoding means in the encryption means 14 / encryption decryption means 19, and various other related encoding means. It will be noted that the order of the decoding / decoding means can be varied. Thus, the frequency spectrum is reduced and / or encrypted before or after performing the frequency shift step. Persons in the industry determine the order of operations depending on the function of available means and whether they operate at low or high frequencies. However, in any case, the order of decoding means / operations is preferably the opposite of the order of encoding means / operations. A simple method of demodulating the encoded second audio signal 5e can be demodulated by filtering the passband between Fp-Fc and Fc and then multiplying by cos (2nпFp / Fe).

The receiver 3 further has demodulation means (not shown) for demodulating the command signal 35 and can extract a binary command. Returning to the example of coding by pulse width modulation described above, this extraction can be performed by sequentially performing the following steps.
That is, the synthesized speech signal 6 is filtered by a square filter after the pass band between F′p−Δ / 2 and F′p + Δ / 2 is set as Δ = (F′p−Fp) / 2, and is filtered by an integrating filter. Then look for the peak of the mountain, sample it to + Nb / 2-δ (where δ is a safety factor, equal to Nb / 16) when the mountain comes out, and store the bits thus detected in a buffer To do.

  Further, according to the preferred embodiment shown in FIG. 4, the receiver 3 has gain adaptive control means 23. The gain adaptive control means 23 receives the volume of the second audio signal 5 reproduced by the second audio reproduction means 9 by applying it to the second audio signal 5 from the decoding means 18 by a known gain adaptive control method. It is not related to the magnitude of the signal received by the means 16. In this way, the volume change due to the change in the distance between the transmitter 2 and the receiver 3 is eliminated, and the second audio signal 5 can be reproduced with a constant volume.

In one advantageous embodiment, in order to compensate for the noise of the room in which the device of the invention is placed, before the reproduction of the second audio signal 5 by the receiving method, the noise reduction step is performed in the following way: be able to.
First, the ambient noise is calculated. This calculation is performed at predetermined intervals based on the noise during silent operation. Here, the silent time is a time when the transmitter does not reproduce the first audio signal and the receiver does not reproduce the second audio signal. In a simple calculation method, parameter noise is added to a voice band to be used to determine blank noise, and then a noise spectrum is determined by a fast Fourier transform method. Applying Ephrim and Malah equations, gain can be calculated using a pre-defined 81 × 81 matrix (-40 dB to +40 dB at 1 dB intervals). At the final stage, the denoising signal is determined by inverse fast Fourier transform. Thereby, noise can be removed from the signal received by the receiver, and the disturbance effect of noise can be suppressed.

  As has been seen so far, the synthesized speech signal 6 can carry the identification code of the intended receiver 3. The receiver 3 is then preferably assigned its own identification code. The receiver 3 has means for extracting an identification code transmitted together with the synthesized speech signal 6. The receiver 3 compares the transmitted identification code with its own identification code, and reproduces the second audio signal 5 only when the two identification codes match.

  The identification code specially labeled with the term << Broadcast >> is known as the audible receiver 3 corresponding to all receivers 3, and when receiving a signal with the code << Broadcast >>, the second audio signal 5 Is reproduced by the receiver 3 whatever its own identification code.

  In a special embodiment, the receiver 3 is associated with a toy. This toy is, for example, a doll.

1 is a full block diagram of a system of the present invention. It is a block diagram of the producer of this invention. It is a block diagram of the transmitter of this invention. It is a receiver block diagram of the present invention. The frequency spectrum of the 1st voice signal is shown. In the frequency spectrum of the first audio signal, the high frequency portion is removed. The frequency spectrum of the 1st audio | voice signal after removing a high frequency part is shown. The frequency spectrum of the 2nd voice signal is shown. The frequency spectrum of the 2nd audio | voice signal reduced is shown. The frequency spectrum of all the second audio signals (negative part and positive part) is shown. The frequency spectrum of the negative part of the 2nd voice signal is shown. The frequency spectrum of the 2nd audio | voice signal shifted to the high frequency is shown. The frequency spectrum of the synthetic | combination audio | voice signal which the 1st audio | voice signal and the 2nd audio | voice signal couple | bonded is shown. An operation for extracting the second audio signal from the synthesized audio signal is shown. The frequency spectrum of the command signal is shown. The first audio signal shows the frequency spectrum of the synthesized audio signal obtained by combining the second audio signal and the command signal.

Claims (28)

A method of creating a synthesized voice signal (6) from a first voice signal (4) and a second voice signal (5),
A coding step in which the second audio signal (5) is made into a coded second audio signal (5e) that can be reproduced in a manner inaudible to the human ear (7) by the first audio reproduction means (8);
The first audio signal (4) and the second audio signal (4) are encoded in a manner in which the first audio signal (4) is heard by the human ear (7) and the second audio signal (5) is not heard by the human ear (7). A method for generating a synthesized speech signal, comprising: a speech combining step for combining the speech signal (5e) into a synthesized speech signal (6) that can be reproduced by the first speech reproduction means (8).   The coding step comprises a frequency shift step of shifting the frequency of the second audio signal (5) to a high frequency that cannot be heard by humans in the reproduction frequency spectrum of the first audio reproduction means (8). The synthetic speech signal creation method according to 1.   The method according to claim 1 or 2, wherein the encoding step includes a band reduction step of the second audio signal (5).   4. The method according to claim 1, wherein the encoding step includes an encryption step.   5. A method according to claim 4, wherein the step of encrypting comprises removing the positive part (5d) of the frequency spectrum of the second speech signal (5) and leaving only the negative part.   6. The method according to claim 1, wherein the voice combining step combines the first voice signal (4) and the coded second voice signal (5 e). 7.   The voice combining step applies a first gain to the first voice signal before combining the first voice signal (4) and the coded second voice signal (5e), and the coded second voice signal. The method according to claim 1, wherein a second gain is given to (5e).   The method according to claim 7, wherein the second gain is lower than the first gain.   2. The voice combining step includes the step of removing the high frequency portion (4c) of the frequency spectrum of the first audio signal (4) before combining and retaining only the low frequency portion (4d). The synthetic | combination audio | voice signal preparation method of any one of thru | or 8.   10. The synthesized speech signal generation method according to claim 1, wherein the speech combining step further includes a combining step of adding a command signal (35) at a frequency Fcmd.   11. Synthesis according to claim 10, characterized in that the frequency Fcmd has a higher wavelength than the coded second audio signal (5e), a frequency lower than 15625 Hz for the PAL signal and lower than 15734 Hz for the NTSC signal. Audio signal creation method.   12. A method according to claim 10 or 11, wherein the command signal (35) is a binary command coded by pulse width modulation at 31.25 bits / second.   A synthesized speech signal (6) obtained by the synthesized speech signal creation method according to any one of claims 1 to 9.   Recording medium (24) of a synthesized speech signal (6), characterized in that it is according to claim 10.   Producer (1), characterized in that it produces a synthesized speech signal (6) according to claim 13. A method of reproducing the first audio signal (4) and wirelessly transmitting the second audio signal (5) and the command signal (35),
Generation of the synthesized voice signal (6) according to claim 12 from the first voice signal (4), the second voice signal (5), and possibly the command signal (35), and the synthesis by the first voice reproduction means (8). A voice reproduction and voice transmission method comprising: voice reproduction of an audio signal (6). A method for receiving a synthesized speech signal (6) according to claim 13, comprising:
-Receive the synthesized voice signal (6)
-Extracting the encoded second audio signal (5e) from the synthesized audio signal (6),
-Decoding the encoded second audio signal (5e) and reconstructing the second audio signal;
An audio receiving method comprising the step of reproducing the second audio signal (5); Furthermore, the command signal (35) is extracted from the synthesized speech signal (6),
18. The method according to claim 17, further comprising the step of demodulating the command signal (35) to reconstruct a binary command.   18. The decoding step or the demodulating step has a step symmetric with respect to a coding or modulation step in the synthetic speech signal creation method according to claim 1. Or the audio | voice receiving method of 18.   The decryption means has a << mirror >> transform that reconstructs the positive part (5d) of the second audio signal (5) from the negative part (5c) of the frequency spectrum of the second audio signal (5). The voice receiving method according to any one of claims 17 to 19.   21. The audio receiving method according to claim 17, wherein a gain / gain control step is applied to the second audio signal (5) before the audio reproduction step.   22. The audio receiving method according to any one of claims 17 to 21, further comprising a noise removing step of the second audio signal (5) from the estimation of the noise spectrum before reproduction.   The voice receiving method according to claim 22, wherein the noise spectrum is regularly measured in a quiet period or treated as blank noise.   The receiver (3) has its own identification code, and an identification code for identifying at least the intended receiver (3) is extracted from the synthesized speech signal (6), and the identification code and the synthesized speech signal (6) are identified. 24. Audio reception according to any one of claims 17 to 23, characterized in that the receiver (3) reproduces the second audio signal (5) only when the identification code extracted from is compatible. Method.   Audio reception according to claim 22, characterized in that the identification code << broadcast >> corresponds to all receivers (3), the identification code << broadcast >> is extracted and the second audio signal (5) is reproduced. Method.   Receiver (3), characterized in that it receives a synthesized speech signal (6) according to claim 13 and can implement the speech reception method according to any one of claims 17 to 25.   A toy comprising a receiver (3) according to claim 26. A reproduction system for a first audio signal (4) and a second audio signal (5),
The producer (1) of claim 15, generating a synthesized speech signal (6) according to claim 13 and transmitting the synthesized speech signal (6);
A transmitter (2) for reproducing the first audio signal (4) and transmitting the second audio signal (5) and the command signal (35) wirelessly, or a first audio player,
27. The receiver (3) of claim 26, or the second audio player, wherein the second audio signal (5) is extracted, decoded and reproduced, and simultaneously the command signal (35) is extracted and modulated;
An audio signal reproduction system comprising:

Images